Lowering costs by driving high utilization across deep learning workloads is a crucial lever for cloud providers. We present Singularity, Microsoft's globally distributed scheduling service for highly-efficient and reliable execution of deep learning training and inference workloads. At the heart of Singularity is a novel, workload-aware scheduler that can transparently preempt and elastically scale deep learning workloads to drive high utilization without impacting their correctness or performance, across a global fleet of AI accelerators (e.g., GPUs, FPGAs). All jobs in Singularity are preemptable, migratable, and dynamically resizable (elastic) by default: a live job can be dynamically and transparently (a) preempted and migrated to a different set of nodes, cluster, data center or a region and resumed exactly from the point where the execution was preempted, and (b) resized (i.e., elastically scaled-up/down) on a varying set of accelerators of a given type. Our mechanisms are transparent in that they do not require the user to make any changes to their code or require using any custom libraries that may limit flexibility. Additionally, our approach significantly improves the reliability of deep learning workloads. We show that the resulting efficiency and reliability gains with Singularity are achieved with negligible impact on the steady-state performance. Finally, our design approach is agnostic of DNN architectures and handles a variety of parallelism strategies (e.g., data/pipeline/model parallelism).
@article{arxiv.2202.07848,
title = {Singularity: Planet-Scale, Preemptive and Elastic Scheduling of AI Workloads},
author = {Dharma Shukla and Muthian Sivathanu and Srinidhi Viswanatha and Bhargav Gulavani and Rimma Nehme and Amey Agrawal and Chen Chen and Nipun Kwatra and Ramachandran Ramjee and Pankaj Sharma and Atul Katiyar and Vipul Modi and Vaibhav Sharma and Abhishek Singh and Shreshth Singhal and Kaustubh Welankar and Lu Xun and Ravi Anupindi and Karthik Elangovan and Hasibur Rahman and Zhou Lin and Rahul Seetharaman and Cheng Xu and Eddie Ailijiang and Suresh Krishnappa and Mark Russinovich},
journal= {arXiv preprint arXiv:2202.07848},
year = {2022}
}